The Determinants of Online Review Informativeness: Evidence from Field Experiments on Airbnb
53 Pages Posted: 27 Jun 2018
Date Written: April 20, 2018
Reputation systems are used by most digital marketplaces, but their design varies greatly. We study how reputation system design affects the extent to which ratings and reviews aggregate information by using two experiments and complimentary observational analysis conducted on Airbnb. The first treatment that we study offered guests a $25 coupon in exchange for submitting a review. The second treatment implemented a simultaneous reveal review system, which eliminated strategic considerations from reviews. We show how both experiments made the reputation system more informative and use our findings to quantify the relative importance of mechanisms that cause inefficiency in reputation systems. We find that the largest source of inefficiency in reputation systems is sorting into reviews, whereby those who submit reviews typically have better experiences than those who do not. We also document retaliation and strategic reciprocity in the control group of our experiments but find that these mechanisms had small effects on the rating distribution. Lastly, we use observational analyses to document that social considerations also cause information loss in reputation systems. In summary, we find that reviews are typically informative but that negative experiences are underreported. We then discuss the implications of our findings for reputation system design.
Keywords: Reputation, Market Design, Digital Marketplaces, Field Experiments, Reciprocity, Information, Product Quality
JEL Classification: C93, D47, L14, L86, D8, L18, L83, M31
Suggested Citation: Suggested Citation